MIMO Wiener Model Identification using Radial Basis Functions Neural Networks

S.S.A. Ali and H.N.Al-Duwaish (Saudi Arabia)


Wiener Model, MIMO, RBFNN, LMS


This paper presents a novel technique for the identification of the nonlinear multi-input multi-output (MIMO) Wiener Model, consisting of linear dynamics in cascade with static nonlinearities. The ARMA/RBFNN structure presented in [1] is exteneded for MIMO case. The proposed algorithm estimates the weights of the RBFNN and the coefficients of ARMA model based on least mean squares (LMS). The identification of both linear and nonlinear parts is done si multaneously as compared to the other indirect approaches.

Important Links:

Go Back